Why size may not matter in factor investing

One analysis questions a school of thought that roots for market underdogs

Why size may not matter in factor investing

Investors may be tempted to embrace the story of the underdog, particularly in the context of small-cap investments that promise outsized returns for taking on extreme risks. But a closer examination of the underlying factor for the strategy hints at a grossly exaggerated success story. 

In a piece published by the CFA Institute, FactorResearch Managing Director Nicolas Rabener noted that based on data from the Kenneth R. French Data Library going back to 1926, the US stock market’s bottom decile of companies by market capitalization exhibits annualized volatility of 15.3%, compared to 14.1% for the largest 10%.

While investors would expect a premium for taking on riskier holdings, Rabener noted that the size-factor approach of buying small-cap stocks and shorting large-caps has yielded almost flat returns over the last 90 years or so, with a few notable boom-and-bust cycles.

“The first influential paper on excess Size factor returns was published by Rolf W. Banz in 1981 — the approximate performance peak for the Size factor,” he said. Data showed positive returns across the entire observation period, though they were not particularly robust especially considering the context of decade-long drawdowns. “Making matters worse, the data excludes transaction costs and thus overstates returns.”

Hypothesizing that the size factor’s returns could be revealed through a particular lens of measurement, Rabener and others looked at its record within the US stock market since the year 2000 based on four alternative metrics. Market capitalization and average daily value traded (ADV) showed nearly identical patterns of returns; “the stocks with the largest market capitalizations tend to be among the most traded,” he observed. Weighting companies based on their total assets and total sales, he added, generated negative returns.

The analysis of the alternative size metrics continued with a look at current median market capitalizations for both long and short portfolios. Portfolio characteristics, Rabener and his colleagues found, were comparable across metrics except for total assets and total sales, where companies in the long portfolio as measured by market capitalization featured smaller companies. “This likely reflects a preference among investors for fast-growing, asset-light companies over more mature businesses with greater assets and sales,” he said.

A sector breakdown of the long portfolio of small caps for the different metrics from 2000 to 2018 also showed mixed results. The market cap-based portfolio showed the most diversity across sectors, while enterprise value was dominated by tech-sector stocks. Financials accounted for most stocks in the ADV-based long portfolio, though Rabener said it’s “difficult to explain why”; the portfolios built on criteria of total assets and total sales, meanwhile, were replete with tech and health care stocks.

“[That’s] probably because such companies have few assets … and tend to grow quickly, but often have little or no sales,” Rabener theorized.

But applying the same strategies in different regions, Rabener found, uncovers positive and relatively consistent returns for portfolios with European stock-market exposure since 2000. Alternative metrics also showed somewhat similar trends over time in Europe, as opposed to the US context where the different measures resulted in divergent results.

“Investors were rewarded for buying smaller, riskier companies in Europe, but not in the United States or Japan,” he said. “This would seem to cast doubt on the very existence of the small-cap premium.”


Follow WP on Facebook, LinkedIn and Twitter